U.S. patent application number 11/912343 was filed with the patent office on 2008-11-06 for projection system and method for operating a projection system.
This patent application is currently assigned to KONINKLIJKE PHILIPS ELECTRONICS, N.V.. Invention is credited to Carsten Deppe, Hans Feijen, Sander Habets, Christofher Daniel Charles Hooijer, Tom Munters.
Application Number | 20080273179 11/912343 |
Document ID | / |
Family ID | 37215134 |
Filed Date | 2008-11-06 |
United States Patent
Application |
20080273179 |
Kind Code |
A1 |
Deppe; Carsten ; et
al. |
November 6, 2008 |
Projection System and Method for Operating a Projection System
Abstract
A method for operating a projection system (1) is described,
wherein a brightness level (B1, B2, B3, B4, B5, B6) in an image
(IM) is represented by the total length of a number of light
switch-on phases within a particular image cycle (VT). A high
pressure discharge lamp (2) of the projection system (1) is
operated via a lamp driver with an essentially square-wave
alternating current (IL) SO that an overshoot sequence (O), which
occurs in the alternating current (IL) after a current zero
crossing during operation of the high pressure discharge lamp (2),
has a frequency so high that for each possible brightness level at
least one full period (P) of the overshoot sequence (O) lies
essentially within a light switch-on or switch-off phase (ts)
following the zero crossing. A projection system (1) and a lamp
driver (10) for a high pressure discharge lamp (2) in such a
projection system (1) are also described.
Inventors: |
Deppe; Carsten; (Aachen,
DE) ; Munters; Tom; (Hasselt, BE) ; Hooijer;
Christofher Daniel Charles; (Helmond, NL) ; Feijen;
Hans; (Eindhoven, NL) ; Habets; Sander;
(Berghem, NL) |
Correspondence
Address: |
PHILIPS INTELLECTUAL PROPERTY & STANDARDS
P.O. BOX 3001
BRIARCLIFF MANOR
NY
10510
US
|
Assignee: |
KONINKLIJKE PHILIPS ELECTRONICS,
N.V.
EINDHOVEN
NL
|
Family ID: |
37215134 |
Appl. No.: |
11/912343 |
Filed: |
April 21, 2006 |
PCT Filed: |
April 21, 2006 |
PCT NO: |
PCT/IB06/51240 |
371 Date: |
October 23, 2007 |
Current U.S.
Class: |
353/85 ;
348/E9.027; 348/E9.037; 353/121 |
Current CPC
Class: |
H04N 9/3155 20130101;
H04N 9/3114 20130101; H04N 9/64 20130101; H05B 41/2887 20130101;
Y02B 20/202 20130101; H05B 41/2883 20130101; Y02B 20/00
20130101 |
Class at
Publication: |
353/85 ;
353/121 |
International
Class: |
G03B 21/20 20060101
G03B021/20 |
Foreign Application Data
Date |
Code |
Application Number |
Apr 28, 2005 |
EP |
05103512.9 |
Claims
1-9. (canceled)
10. A method for operating a projection system (1), wherein a
brightness level (B.sub.1, B.sub.2, B.sub.3, B.sub.4, B.sub.5,
B.sub.6) in an image (IM) is represented by the total length of a
number of light switch-on phases within a particular image cycle
(VT), and wherein a high pressure discharge lamp (2) of the
projection system (1) is operated via a lamp driver with an
essentially square-wave alternating current (I.sub.L) so that an
overshoot sequence (O), which occurs in the alternating current
(I.sub.L) after a current zero crossing during operation of the
high pressure discharge lamp (2), has a frequency so high that for
each possible brightness level at least one full period (P) of the
overshoot sequence (O) lies essentially within a light switch-on or
switch-off phase (t.sub.S) following the zero crossing.
11. A method as claimed in claim 10, characterized in that the high
pressure discharge lamp (2) is operated so that the frequency of
the overshoot sequence (O) is greater than or equal to 7.5 kHz.
12. A method as claimed in claim 11, characterized in that the high
pressure discharge lamp (2) is operated so that the frequency of
the overshoot sequence (O) is greater than or equal to 9.5 kHz.
13. A method as claimed in claim 10, characterized in that a
brightness level (B.sub.1, B.sub.2, B.sub.3, B.sub.4, B.sub.5,
B.sub.6) in the image (IM) is represented digitally by a sequence
of bits with different values within an image cycle (VT), the bits
respectively being defined by the length of a light switch-on phase
in the relevant image cycle (VT), and the high pressure discharge
lamp (2) being operated so that the frequency of the overshoot
sequence (O) is so high that at least one full period (P) of the
overshoot sequence (O) lies essentially within the switching phase
of a least significant bit.
14. A method as claimed in claim 10, characterized in that a
current pulse (I.sub.AF) having the same polarity as the operating
current (I.sub.L) is superimposed on the operating current
(I.sub.L) before a current zero crossing.
15. A method as claimed in claim 14, characterized in that the
projection system (1) comprises a color filter (4) for time
sequential color representation, which alternates between different
colors (r, g, b) during an image cycle (VT), and the color change
is synchronized with the alternating current (I.sub.L) so that the
current pulse (I.sub.AF) occurs within a time span (s) required for
a color change.
16. A projection system (1) having a high pressure discharge lamp
(2), a display device (6) with a control device (9) which switches
the display device (6) in a time-modulated fashion so that a
brightness level (B.sub.1, B.sub.2, B.sub.3, B.sub.4, B.sub.5,
B.sub.6) in an image (IM) is represented by the total length of a
number of light switch-on phases within a particular image cycle
(VT), and a lamp driver (10) which is designed such that the high
pressure discharge lamp (2) of the projection system (1) is
operated with an essentially square-wave alternating current
(I.sub.L) so that an overshoot sequence (O), which occurs in the
alternating current (I.sub.L) after a current zero crossing during
operation of the high pressure discharge lamp (2), has a frequency
so high that for each possible brightness level (B.sub.1, B.sub.2,
B.sub.3, B.sub.4, B.sub.5, B.sub.6) at least one full period (P) of
the overshoot sequence (O) lies essentially within a light
switch-on or switch-off phase (t.sub.S) following the zero
crossing.
17. A lamp driver (10) for a high pressure discharge lamp (2) in a
projection system (1), having voltage source terminals (16) for
connecting a DC voltage source (V1), lamp terminals (21) for
connecting the high pressure discharge lamp (2), a circuit
arrangement which converts a direct current tapped from the DC
voltage source (V1) into an essentially square-wave alternating
current (I.sub.L) for the high pressure discharge lamp (2) with a
frequency of between 40 Hz and 2 kHz, preferably between 45 and 150
Hz, the circuit arrangement being designed so that an overshoot
sequence (O), which occurs in the alternating current (I.sub.L)
after a current zero crossing during operation of the high pressure
discharge lamp (2), has a frequency greater than or equal to 7.5
kHz, preferably greater than or equal to 9.5 kHz.
18. A lamp driver as claimed in claim 17, characterized in that the
circuit arrangement has a switching converter (17), which sets the
amplitude of the operating current of the high pressure discharge
lamp (2), and a commutator circuit (18), which commutates the
current direction, wherein essentially inductances (LT1, LT2) in
the commutator circuit (18) and capacitances (C1) in the switching
converter (17) are being dimensioned so that the frequency of the
overshoot sequence (0) is greater than or equal to 7.5 kHz,
preferably greater than or equal to 9.5 kHz.
Description
[0001] The invention relates to a projection system having a high
pressure discharge lamp, preferably a high performance discharge
lamp, for example a UHP (ultra-high performance) lamp, and a
display device with a control device which switches the display
device in a time-modulated fashion so that a brightness level in an
image is represented by the total length of a number of light
switch-on phases within a particular image cycle. The invention
also relates to a method for operating a projection system, and to
a lamp driver for a high pressure discharge lamp in such a
projection system.
[0002] A typical example of such a projection system is the
so-called DLP.RTM. system (DLP=digital light processing) from Texas
Instruments.RTM.. The key component of this system is a display
device in the form of a chip, on which a number of minute mobile
mirrors is applied as individual display elements--one mirror per
image point to be represented. These mirrors are exposed to the
light from the high pressure discharge lamp. Depending on whether
an image point on the projection surface, i.e. in the image to be
represented, is intended to appear bright or dark, the associated
mirror is tilted so that the light is reflected toward the
projection surface or away from it to an absorber. The individual
mirrors of this "mirror array" consequently form a network with
which arbitrary image patterns can be generated and, for example,
video images can be reproduced. Since the light at an image point
of the projection surface can only be turned on or off in such a
system by tilting the associated mirror--and defined attenuation of
the light cannot be achieved, for example as in LCD (Liquid Crystal
Display) systems--the various brightness levels in the image (or
gray levels in a monochrome image) are represented in the way
mentioned in the introduction by means of a pulse width modulation
method (also known as a "time sequential" method). Within a
particular image cycle, each display element of the display device
is driven so that, overall, light strikes the relevant image point
of the projection surface over a particular time fraction of the
image cycle, and not in the remaining time fraction. Owing to the
inertia of the eye relative to the image cycle frequency, the
amount of light is integrated in an observer's eye over an image
cycle, so that the relevant image point appears brighter or darker.
If the light is turned off fully during an image cycle, i.e. a
light switch-off phase occupies the entire image cycle then the
relevant image point appears black. If the light is turned on over
the entire length of the image cycle, however, then the relevant
pixel has the highest brightness level. Since the brightness in
such systems is controlled via the time sequential switching, i.e.
the length and number of the light switch-on and switch-off phases
within an image cycle, it is necessary that the light intensity to
which the display is exposed should be as constant as possible not
only spatially but also temporally. This is because if the light
intensity were to change within an image cycle, then the brightness
level would be determined not only by the total length of the light
switch-on phases and light switch-off phases but also by the
chronological position of the individual light switch-on phases
within the image cycle.
[0003] For illuminating the display element, it is consequently
necessary to have a light source which as far as possible emits a
constant luminous power (a constant light flux), i.e. always the
same amount of light. It is now conventional to use high pressure
discharge lamps as lamps in such projection systems, preferably
high intensity discharge lamps (HID lamps). In order to ensure
uniform wear of the two electrodes and therefore a sufficient
service life of the lamp, these lamps are operated with a
low-frequency square-wave alternating current so that the current
direction between the electrodes switches over constantly. The lamp
is in this case generally driven by a special lamp driver, which
converts the DC current provided by a DC voltage source into the
square-wave alternating current required for the lamp with the
appropriate pulse shape. One disadvantage of this method is that
unless particular measures are implemented, the abrupt switchover
of the current direction in conventional drivers necessarily leads
to an overshoot sequence in the operating current of the lamp after
the zero crossing. This overshoot sequence is also referred to as
ringing. The overshoot sequence in the operating current of the
lamp naturally needs to a corresponding overshoot in the luminous
power emitted by the high pressure discharge lamp. On the other
hand--as mentioned above--it is necessary for the lamp to have a
constant luminous power. In order to resolve this problem, it would
in principle be possible to use suitable synchronization in order
to ensure that the overshoot sequence lies outside an image cycle.
In many projection systems, for example, the color representation
is likewise carried out time-sequentially. To this end, three
monochrome images are generated in succession--a red image, a green
image and a blue image. The inertia of the eye then ensures that
the image appears colored. To this end, for instance, a color
filter is arranged between the lamp and the display device, usually
in the form of a so-called color wheel which alternates the color
at a particular frequency. In the transition times from one color
to another--referred to as spokes in analogy with the spokes of a
wheel--the color is undefined. These spoke times are therefore
inherently unusable for the normal image representation in the
image, and are either blocked out or used to reinforce the
brightness of the display with a reduced quality. The square-wave
alternating current could therefore in principle be synchronized
with the driving of the color filter so that the overshoot sequence
always lies within such a spoke time. On the other hand, as
described in U.S. Pat. No. 5,608,294, for the stability of the arc
discharge and the durability of the electrodes of a high pressure
discharge lamp it is advantageous that a current pulse, which
temporarily increases the operating current before the zero
crossing, should be superimposed on the operating current as close
as possible before a current zero crossing. Such a stabilization
pulse is also generally referred to as an anti-flatter pulse. Such
an anti-flatter pulse of course also leads to corresponding light
fluctuations in the lamp. So that such an advantageous operating
method can also be used in the time-sequentially operating
projection systems mentioned in the introduction, EP 1 154 652 A2
proposes that the lamp driver and the controlling of the color
filter should be synchronized so that this anti-flatter pulse lies
within a spoke time and cannot lead to disturbance of the
brightness levels in the image. Since the anti-flatter pulse lies
immediately before the zero crossing of the current and the
overshoot sequence lies immediately after the zero crossing, these
regions directly follow each other inside the time profile of the
current. Unfortunately, however, the spoke time is generally at
best long enough for the anti-flatter pulse to be accommodated in
this time span. The overshoot sequence therefore necessarily comes
after a spoke time, i.e. at the start of an image cycle. In
principle, it would admittedly be possible to extend the spoke time
artificially by appropriate driving of the display. This solution,
however, leads to a great loss of light which degrades the overall
image quality.
[0004] Conventionally, this problem is therefore countered by
designing the driving of the display elements so that the system
reacts as insensitively as possible to the overshoot sequence. For
example, the individual switch-on or switch-off times within an
image cycle may be selected so that as long as possible a
switch-off phase lies directly at the start of an image cycle and
the overshoot sequence is therefore substantially blocked out.
Although this avoids the effect of the overshoot sequence on the
brightness value representation, nevertheless there are then still
very stringent requirements on the driver in order to suppress the
visible influences of the overshoot sequence.
[0005] It is an object of the present invention to refine a method
for operating a projection system, and a projection system, of the
type mentioned in the introduction so that the influence on the
representation of the brightness levels in the image due to the
overshoot sequence, which occurs in the operating current of the
high pressure discharge lamp after a current zero crossing, is
reduced as far as possible in a straightforward and cost-effective
way.
[0006] This object is achieved, on the one hand, by a method as
claimed in claim 1.
[0007] According to the invention, the high pressure discharge lamp
of the projection system is operated via a lamp driver with an
essentially square-wave (low-frequency) alternating current so that
the overshoot sequence, which occurs in the alternating current
after a current zero crossing has a frequency so high that for each
possible brightness level at least one full period of the overshoot
sequence lies essentially within a light switch-on or switch-off
phase following the zero crossing. To resolve the problem, an
attempt is not made--as would otherwise be customary--to minimize
the amplitude of the overshoot or even entirely avoid the overshoot
sequence, but instead the frequency is simply increased so that the
effects of the overshoot sequence are avoided because at least one
full period lies within a switch-on or switch-off phase following
the zero crossing. This effect can be further enhanced by suitable
time positioning or length selection of the individual light
switch-on and switch-off phases. In a plurality of test runs, it
has been found that a frequency increase according to the invention
can achieve a significant improvement in the brightness level
representation, and that above particular frequencies absolutely no
effect of the overshoot sequence on the brightness level
representation can be discerned by the observer. The requirements
of the lamp driver with respect to the amplitude of the overshoots
can therefore be relaxed.
[0008] The term an "essentially square-wave alternating current" in
the context of this invention moreover generally means an
alternating current with a current profile which--apart from
customary tolerances and current pulses deliberately superimposed
for particular effects--has a substantially constant profile over
time and a sufficiently fast polarity change. That is to say
besides a customarily used square-wave alternating current, this
also includes for example a trapezoidal alternating current or the
like, the edges of which are steep enough for use in a
corresponding projection system. The current pulses deliberately
superimposed for particular effects may, for example, be the
aforementioned anti-flatter pulse or in particular square-wave
pulses which, for example, extend over the length of a particular
color segment in order to increase the intensity of this color
compared to other colors.
[0009] The object is furthermore achieved by a projection system as
claimed in claim 7.
[0010] Besides a high pressure discharge lamp and a conventional
display device with a control device, which switches the display
device in a time-modulated fashion so that a brightness level in an
image is represented by the total length of a number of light
switch-on phases within a particular image cycle, such a projection
system according to the invention requires above all a lamp driver
which is designed such that the high pressure discharge lamp of the
projection system is operated with an essentially square-wave
alternating current so that the overshoot sequence, which occurs in
the alternating current after a current zero crossing during
operation of the high pressure discharge lamp, has a frequency so
high that for each possible brightness level at least one full
period of the overshoot sequence lies essentially within a light
switch-on or switch-off phase following the zero crossing.
[0011] Preferably, the frequency should be so high that if possible
a multiplicity of full periods of the overshoot sequence lies
within a switch-on or switch-off phase. In principle, however, in
the extreme case it is sufficient for such averaging out of the
effect of the overshoot sequence that one full period should lie
within such a switch-on or switch-off phase.
[0012] The dependent claims respectively contain particularly
advantageous configurations and refinements of the invention.
[0013] In a plurality of tests, it has been found that relatively
good results are achieved if the frequency of the overshoot
sequence is greater than or equal to 7.5 kHz, so that it is
preferable to ensure that the frequency reaches this value.
Particularly preferably, a frequency which is greater than or equal
to 9.5 kHz should be reached, since the effect of the overshoot
sequence can no longer be discerned by the observer above this
frequency in the systems tested so far.
[0014] In the aforementioned projection systems, a brightness level
in the image is conventionally represented digitally by a sequence
of bits with different values within an image cycle. The bits are
respectively defined by the length of a light switching phase in
the relevant image cycle. In principle, the bits may also be
decomposed into a plurality of switching sub-phases. In order to
ensure that for each possible brightness level, i.e. even with the
lowest brightness, the effect of the overshoot sequence is minimal
irrespective of how the sequence of individual bits--i.e. their
individual switching phases--is arranged or distributed within an
image cycle, the frequency in a more particularly preferred
embodiment should be selected so that at least one full period of
the overshoot sequence lies essentially within the switching phase
of a least significant bit (conventionally called LSB).
[0015] Besides voltage source terminals for connecting a DC voltage
source and lamp terminals for connecting the high pressure
discharge lamp, a particularly preferred lamp driver constructed
according to the invention for a high pressure discharge lamp in a
projection system according to the invention has a circuit
arrangement which converts a direct current tapped from the DC
voltage source into an essentially square-wave alternating current
for the high pressure discharge lamp with a frequency of between 40
Hz and 2 kHz, preferably between 45 and 150 Hz. The circuit
arrangement is in this case designed so that an overshoot sequence,
which occurs in the alternating current after a current zero
crossing during operation of the high pressure discharge lamp, has
a frequency greater than or equal to 7.5 kHz, preferably greater
than or equal to 9.5 kHz.
[0016] If the circuit arrangement has a switching converter which
sets the amplitude of the operating current of the high pressure
discharge lamp, i.e. it imposes the current profile, and a
commutator circuit which commutates the current direction, then
essentially the inductances in the commutator circuit and
capacitances in the switching converter are preferably dimensioned
so that the frequency of the overshoot sequence corresponds to the
aforementioned values. When dimensioning the inductances inside the
commutator circuit, the inductance of the lamp itself should also
be taken into account so that the inductances acting in the
commutator circuit must be designed in accordance with the other
components.
[0017] In the scope of the invention, not only is the frequency of
overshoots occurring after a zero crossing increased but preferably
also the frequency of the overshoots which may occur after other
steep current changes, for example after the aforementioned current
pulses deliberately superimposed on a "basic alternating current".
This effect is advantageously achieved without additional outlay in
the aforementioned configuration of the lamp driver.
[0018] These and other aspects of the invention are apparent from
and will be elucidated with reference to the embodiments described
hereinafter.
[0019] In the drawings:
[0020] FIG. 1 shows a schematic representation of an exemplary
embodiment of a projection system,
[0021] FIG. 2 shows a schematic representation to explain the
time-sequential representation of brightness levels within an
image,
[0022] FIG. 3 shows a schematic representation of the lamp
operating current, the resulting light flux and the synchronization
clock for driving various components of the projection system,
[0023] FIG. 4 shows a diagram which represents the relative error
in the light flux against the frequency of the overshoot sequence
for various switching times,
[0024] FIG. 5 shows a basic circuit diagram of an exemplary
embodiment of a lamp driver according to the invention.
[0025] FIG. 1 shows a typical DLP.RTM. projection system 1. This
system has a high pressure discharge lamp, preferably a so-called
UHP (ultra-high performance) lamp, arranged in a reflector 3. The
lamp 2 is supplied with the required square-wave voltage by a
driver 10. The light emitted by the lamp 2 is focused in the
direction of a converging lens 5 by the reflector 3. In the beam
path of the light, there is a color wheel 4 between the reflector 3
and the converging lens 5. The color wheel 4 is preferably
positioned as accurately as possible at the focal point lying
between the reflector 3 and the converging lens 5 so that the light
spot on the color wheel 4 is as small as possible. The converging
lens 5 ensures that a display device, here a so-called DMD (digital
mirror device), located behind the converging lens 5 is illuminated
as well as possible. As already mentioned in the introduction, such
a DMD is a chip with a matrix of minute mirrors 7 as display
elements 7, which can be tilted individually so that the mirrors 7
reflect the light either through a projection objective 8 onto a
projection surface F (see FIG. 2) or, with an appropriately tilted
mirror 7, onto an absorber surface (not shown).
[0026] The color wheel 4 here has three segments with the colors
red r, green g and blue b. This color wheel 4 rotates one or more
times about its own axis within an image cycle VT, so as to
successively generate a monochrome blue image, a monochrome red
image and a monochrome green image which combine to form a color
image owing to the inertia of the observer's eye. Instead of a
color wheel with three segments, it is of course also possible to
use other color wheels, for example color wheels with four segments
(a red segment, a green segment, a blue segment and a white
segment) or with six segments, in which two green, two red and two
blue segments respectively lie opposite each other. The latter
variant has the advantage that the color wheel only has to rotate
with half the frequency. Alternatively, further different color
segments may also be introduced, for example yellow, cyan, etc.
[0027] A control device 9 drives the display device 6 (also
referred to below as the "display" for brevity), the color wheel 4
and the lamp driver 10 and thus ensures the synchronization, to be
described in more detail below, of these components 6, 4, 10. As
the input signal, for example, the control device 9 may receive a
video signal V which contains video data that are to be represented
by the projection system 1. It is clear that the control device 9
may also consist of a plurality of control sub-devices. For
example, a separate control device may also be provided in order to
drive the display elements 7 of the display 6, and this may in
particular also be arranged on a circuit board with the DMD chip.
It is likewise possible for the complete control device to be
integrated in one component, for example with the DMD chip, and to
send the synchronization commands for the other components from
there. It is furthermore possible for the other components, such as
the color wheel or lamp driver, to have their own control devices
which in turn transmit synchronization commands to the other
components. What is essential in the end is merely that the
components are synchronized with one another in the required way so
that the intended image is generated on the projection surface
F.
[0028] FIG. 2 shows a graphical representation to explain how
different brightness levels B.sub.1, B.sub.2, B.sub.3, B.sub.4,
B.sub.5, B.sub.6 can be represented in the image IM by
time-modulated switching of the individual mirrors 7. A DMD chip
with 6.times.6 individual mirrors 7 is represented here by way of
example. By the projection objective 8, the light rays reflected by
the individual mirrors 7 are represented as image pixels within the
image IM on the projection surface F. For the sake of clarity, in
this example only six different brightness levels B.sub.1, B.sub.2,
B.sub.3, B.sub.4, B.sub.5, B.sub.6 are shown in one of the two
central columns of the matricial image IM.
[0029] The time profile within an image cycle VT is represented
between the projection objective 8 and the projection surface F.
The example represented is based on a simplified version in which
16 different brightness levels can be represented digitally with
the aid of four bits. Technically, this is done by respectively
defining the bits by the total length of a particular light
switching phase in the relevant image cycle VT, i.e. the value of a
bit depends on how long the light strikes the respective image
pixel on the projection surface F during an image cycle VT, i.e. is
switched on. Examples of bit sequences, which may be set from top
to bottom for the individual mirrors 7 in the central column of the
DMD chip 6, are shown below the DMD chip 6 (in FIG. 2). The top
mirror is accordingly driven in an image cycle VT so that the
corresponding image pixel has the lowest brightness level B.sub.6.
This mirror is accordingly driven with the value 0, which
corresponds in binary to the number [0000]. The mirror lying below
is intended to represent the next brightness level up B.sub.5, and
is therefore driven with the binary number [0001]. That is to say
this mirror is switched so that the light shines only during a time
span which corresponds to the least significant bit, and is
otherwise switched off. When there are 16 levels in all, this time
span corresponds to one 16.sup.th of the total image cycle VT,
which is correspondingly represented in FIG. 2. The third mirror
down is intended to represent the next brightness level up and
therefore receives the binary value [0010], to which end this
mirror is switched on for a time period of one 8.sup.th of the
image cycle VT and switched off for the rest of the time. In this
way, 16 brightness levels can be arbitrarily represented digitally,
with the mirror reflecting the light onto the relevant pixel on the
projection plane throughout the image cycle VT in the case of the
highest brightness level B.sub.1. This is represented in FIG. 2 for
the bottom mirror and the associated top image pixel in the
relevant column of the image IM.
[0030] If a signal is now to be represented by discrete digital
values in this way, then the resolution is set by the number of
possible levels. The standard for modern projection systems is an 8
bit representation. This corresponds to 256 brightness levels in
all. In general, it is necessary for the light flux to be so
constant that the accuracy of the representation is at least better
than or equal to the lowest brightness levels. If errors greater
than the lowest representable level occur in the representation,
there may sometimes no longer be constant function. This means that
a value which should actually be minimally brighter is darker
because of possible errors in the consistency of the light flux.
Such errors are unfortunately particularly perceptible in images
since they lead to stripes in slow color profiles.
[0031] An estimate will be given below for the maximum permissible
relative error in the consistency of the light flux which would
still be allowed without leading to visible errors in the image.
For the sake of simplicity, the average light flux which the lamp
emits over time is set as 100% for this. The switch-on time of a
bit is conventionally indicated in .mu.s. The amount of light
within a bit consequently has the unit .mu.s %.
[0032] As mentioned above, the accuracy of the representation must
be at least better than or equal to the lowest representable level.
Assuming a conventional image rate of 60 Hz--three color images
being represented successively within an image cycle--and the
conventional 8 bit representation (256 levels), then the switch-on
time of the least significant bit is
t.sub.LSB=1/(603256)=21.7 .mu.s (1)
With a constant light flux of 100%, this corresponds to an amount
of light of 2,170 .mu.s %.
[0033] All interference which could be caused by errors in the
light flux should consequently be less than 2,170 .mu.s %. If an
overshoot sequence with a duration of 200 .mu.s occurs in the light
flux, then this must on average have an amplitude of at most
2,170 .mu.s %/200 .mu.s=10.85%
[0034] of the average amplitude of the light flux.
[0035] In this regard, FIG. 3 shows a schematic representation of
the curve profiles of the various parameters within the projection
system against time during an image cycle VT.
[0036] The curve S shows the profile of the synchronization signal,
which is used to synchronize the lamp driver 10, the color wheel 4
and the display 6 with one another.
[0037] The curve I.sub.L shows the profile of the lamp current. A
square-wave alternating current is employed here, the half-period
of which corresponds to the length of an image cycle VT. This means
that with an image frequency of 60 Hz, for example, the square-wave
alternating current of the lamp driver 10 has a frequency of 30 Hz.
The lamp driver 10 shapes the current so that a current pulse
I.sub.AF, the so-called anti-flatter pulse, is set immediately
before switchover of the current direction. As explained in the
introduction, this ensures that the arc discharges remain as stable
as possible and the durability of the electrodes is extended. The
undesired overshoot sequence O then occurs immediately after the
zero crossing.
[0038] The curve Z shows the profile of the light intensity on the
projection surface F (assuming that the light has not in the
meantime been tilted away with the aid of the mirror).
[0039] Light fluctuations due to the overshoot sequence O in the
current profile can be seen clearly in the light flux profile Z.
This means that the overshoot sequence also occurs almost unchanged
in the light flux profile Z. At the start and end, as well as at
two positions during the image cycle VT, the light flux profile Z
furthermore shows the spoke times s in which the light is blocked
during the color change. The various color phases red r, green g
and blue b lie between spoke times s.
[0040] As the profiles of the curves S, Z, I.sub.L show, the lamp
driver 10 and the color wheel 4 are synchronized so that the
anti-flatter pulse I.sub.AF in the current profile I.sub.L
coincides exactly with a spoke time S. This, however, means that
the overshoot sequence O following the anti-flatter pulse I.sub.AF
lies after the zero crossing inside the image cycle VT. According
to the calculations above, it is therefore necessary to ensure that
the error generated by the overshoot sequence O is as small as
possible. This is done according to the invention by increasing the
frequency of the overshoot sequence O to such an extent that at
least one full period P lies within a coherent switch-on or
switch-off phase t.sub.S of the display following the zero
crossing, as indicated in FIG. 3.
[0041] Here, it should be borne in mind that appropriate
positioning of the individual bits within the image cycle VT can of
course ensure that the coherent switch-on or switch-off phase
t.sub.S at the start of an image cycle, i.e. in the region where
the overshoot sequence occurs, is relatively long. The largest
representable bit, which has a length of 2,560 .mu.s with a mirror
switching time of 20 .mu.s as currently used in general, may for
example always be set at the start of an image cycle VT.
Conventionally, however, the switch-on time of the larger bits as
well is further divided into smaller switching phases, since
excessively long switch-on times can make flickering occur in
moving images or if the eyes are moved over the image. The duration
of the individual switch-on and switch-off phases is therefore
usually of the order of 100 to 1,000 .mu.s in practice, with small
switching phases preferably being used.
[0042] The light flux profile L(t) during the overshoot sequence
can be mathematically described as follows:
L ( t ) := 1 + A ring e - ( t .tau. ring ) sin ( 2 .pi. f ring t -
.pi. 6 ) ( 2 ) ##EQU00001##
Here, f.sub.ring is the frequency, A.sub.ring is the amplitude and
T.sub.ring is the time constant of the decay of the overshoot
sequence.
[0043] The deviation in the light flux .DELTA.L can therefore be
calculated as a function of the frequency f.sub.ring of the
overshoot sequence and according to the length of the switching
phase t.sub.S placed after the zero crossing:
.DELTA. L ( t S , f ring ) := [ [ .intg. 0 t S [ 1 + A ring e - ( t
.tau. ring ) sin ( 2 .pi. f ring t - .pi. 6 ) ] t ] - t S ] 10 6
100 L L S B ( 3 ) ##EQU00002##
L.sub.LSB is in this case the length of the least significant bit.
The function is normalized without units to 1 by the term
10.sup.6100/L.sub.LSB.
[0044] The function given by Equation (3) for the deviation
.DELTA.L is represented in FIG. 4 as a function of the frequency
f.sub.ring of the overshoot sequence for various switching times
TS. The length of the least significant bit is in this case set to
L.sub.LSB=10 .mu.s, which corresponds to the desired standard for a
newer generation of mirror arrays with higher switching times of 10
.mu.s. The amount of light of the least significant bit would then
be 1,000 .mu.s % according to the above calculations for Equation
(1). The curves in FIG. 4 respectively represent the error against
the frequency of the overshoot sequence for switching times t.sub.S
of between 90 .mu.s and 700 .mu.s, with the function corresponding
to
[0045] .DELTA.L.sub.1: t.sub.S=90 .mu.s,
[0046] .DELTA.L.sub.2: t.sub.S=100 .mu.s,
[0047] .DELTA.L.sub.3: t.sub.S=125 .mu.s,
[0048] .DELTA.L.sub.4: t.sub.S=150 .mu.s,
[0049] .DELTA.L.sub.5: t.sub.S=300 .mu.s,
[0050] .DELTA.L.sub.6: t.sub.S=700 .mu.s.
[0051] The curves in FIG. 4 are of course also dependent on the
amplitude of the overshoot sequence. A relatively high value was
assumed to calculate the curves in FIG. 4, for which the amplitude
of the overshoot frequency may be up to 40% of the average "normal"
current amplitude (instead of the values <10% currently adhered
to). The allowed error is less than 1. The curves clearly show that
with a switching time t.sub.S of 100 .mu.s, the error falls below 1
for a frequency above about 7,800 Hz. With the frequency of about
5,000 Hz which is currently conventional for the overshoot
sequence, however, the error is above 1 for all switching times.
This clearly shows that increasing the frequency f.sub.ring of the
overshoot frequency can lead to a significant reduction of the
error in the light flux. For the realistic conditions represented
in FIG. 4, which already take into account a reduced switching time
of 10 .mu.s for the least significant bit, increasing to above 7.8
kHz is sufficient to compensate for the error in the light flux
generated after commutation by the overshoot sequence with a
relatively small switching time t.sub.S of 100 .mu.s.
[0052] FIG. 5 shows a lamp driver 10 through which a corresponding
current profile can be achieved with a sufficiently high frequency
of the overshoot sequence in a straightforward way by suitable
dimensioning of various components. The lamp 2 is connected to this
driver 10 via suitable lamp terminals 21.
[0053] This lamp driver 10 is connected to a DC voltage supply V1
via voltage supply terminals 16. The external DC supply voltage is
for example 380 V. On the input side, the driver 10 has a switching
converter 17 which is responsible for imposing the intended lamp
current. This switching converter 17 is formed by a switch M1, a
diode D1, an inductance L1 and a capacitance C1, for example a
capacitor. A control device 11 controls the switch M1, and
therefore the current in the lamp 2, via a level shifter 13.
Depending on the embodiment, the current may also be monitored by
the control device 11 using an inductive measuring element 14. For
the voltage measurement, a correspondingly reduced voltage is
tapped from the capacitor C1 via a voltage divider 15, consisting
of two resistors R1 and R2, and is measured in the control device
11 by means of an analog-digital converter 12. The capacitor C2 is
used merely to reduce interference in the measurement signal.
[0054] The correspondingly pre-shaped current is then transmitted
to a commutator circuit 18 that comprises a commutator driver 19,
which switches four switches M2, M3, M4, M5 in the manner of a
bridge circuit and thus correspondingly switches over the current
to the lamp 2.
[0055] For ignition, the lamp 2 is furthermore coupled to an
ignition transformer 20. This ignition transformer 20 is
conventionally fitted symmetrically to both terminals 21 of the
lamp 2, as shown here. The ignition transformer 20 provides up to
20 kV for igniting the lamp 2. The inductances LT1, LT2 of the
ignition transformer 20 also act to smooth the current during
further operation of the lamp 2.
[0056] By corresponding model calculations, it can be established
that the frequency f.sub.ring of the overshoot sequence after the
zero crossing is primarily influenced by the sum L of the lamp
inductance, the ignition transformer inductance(s) LT1, LT2 and any
further series inductances present in the commutator circuit 18, on
the one hand, and by the filter capacitor C1 in the switching
converter 17 on the other hand. To a relatively good approximation,
the frequency can then be described as follows:
f ring = 1 2 .pi. L * C 1 ( 4 ) ##EQU00003##
[0057] The inductance L.sub.lamp of the lamp 2 is usually itself
about 300 .mu.H. Ignition transformers are currently used in
conventional lamp drivers, so that the inductance L in the
commutator circuit 18 is about 500 .mu.H. As a rule, a capacitance
C1 of 1 .mu.F has previously been used in the switching converter.
According to Equation (4), this gives 5,627 Hz for the frequency of
the overshoot sequence occurring in the lamp drivers which are
currently conventional.
[0058] If a capacitance C1 of 0.5 .mu.F is instead used in the
switching converter 17, then the frequency of the overshoot
sequence can easily be increased to f.sub.ring=7,958 Hz. If
suitable selection of the ignition transformers is furthermore used
to ensure that the inductance L in the commutator circuit 18 is
halved, then a frequency f.sub.ring=9,597 Hz is in fact reached.
This shows that even in previous driver technology, the desired
frequency increases into a frequency range of more than 7,500 Hz,
preferably 9,500 Hz, are already possible with straightforward
measures and the image quality can be improved significantly.
[0059] To conclude, it should again be pointed out that the systems
and methods represented in the figures and the description are
merely exemplary embodiments which can be varied to a wide extent
by the person skilled in the art, without departing from the scope
of the invention. For the sake of completeness, it will also be
pointed out that the use of the indefinite article "a" or "an" does
not exclude the possibility that there may also be several of the
relevant features.
* * * * *